1. Field of the Invention
The present invention generally relates to an apparatus and method for generating ultrashort optical pulses at a plurality of optical wavelengths, and, more particularly, to an apparatus and method using optical fibers and optical waveguides to produce and control such optical pulses. Ultrashort is here generally referred to as being within the time scale of approximately 10.sup.-15 seconds (femtoseconds) to 10.sup.-12 seconds (picoseconds). The present invention further relates to a method and apparatus for optical imaging using ultrashort optical pulses simultaneously emitted at a plurality of optical wavelengths.
2. Description of the Related Art
A variety of laser systems for producing ultrashort optical pulses is known in the prior art. From a practical point of view, these systems can be generally grouped into two main categories: solid-state laser systems, which are based on the use of volume laser gain media, and fiber laser systems, which are based on waveguiding fiber-optic components. Due to their intrinsic structure, fiber lasers have a number of basic properties which make them significantly more suitable for widespread practical use. As is well known in the prior art, fiber lasers are compact, can be diode pumped, and are robust and reliable. For a number of reasons, at present, the most mature technology suitable for ultrashort-pulse fiber laser systems is based on Er-doped fiber providing output pulses having a wavelength of approximately 1.55 .mu.m. First, Er-doped fibers are among the best developed of-the rare-earth-doped fibers. Diode lasers for pumping such fibers are also well advanced.
Significantly, the generation of ultrashort pulses requires design-control of the dispersion in the laser cavity. This can be accomplished in a compact, all-fiber cavity only at wavelengths above 1.3 .mu.m, where the dispersion of the optical fiber can be tailored to be either of positive or negative sign. However, a variety of practical applications for ultrashort pulses require other wavelengths of operation, for example, either at shorter or longer wavelengths. At those wavelengths, femtosecond-pulse fiber oscillators at present can be designed only by using bulky external components, such as sets of prism pairs, to control the in-cavity dispersion.
For many applications, the wavelength of the laser is critical. For example, for confocal microscopy used in cellular biology, specific dyes are attached to different parts of the cell and are used to observe different functions. Each of these dyes is excited to fluorescence by a respective spectrum of light. Thus, for confocal microscopy, a plurality of different wavelength lasers are used for different dyes. Recently, ultrafast lasers with short pulses and high peak power have been used to excite dyes at resonances which require two-photon excitation. That is, ultrafast lasers have been used to supply enough photon density at the focus of a microscope to cause a non-linear optical effect, called the two-photon absorption effect. This effect is used to excite dyes at the energy level which corresponds to half of the wavelength of each of the two original photons. However, the number of lasers available at different wavelengths is limited; consequently there are only a few dyes which can be utilized at this time. Therefore, the field of two photon microscopy could benefit greatly from a laser capable of being widely tuned to the different wavelengths corresponding to a number of different dyes. The current accepted specifications for a laser for scanning two-photon microscopy are 10-30 mw average power, 100-200 fs pulse width and 50-100 MHZ repetition rate.
The general and well known method to extend the wavelength range of any particular laser system is to utilize nonlinear optical interactions, such as optical harmonic generation, sum or difference frequency generation and optical parametric gain.
Harmonic generation is suitable only for converting an optical signal to a higher optical frequency (shorter wavelength) and it cannot provide tunable or multiple-wavelength output. Sum-frequency and difference-frequency generation allows conversion of a signal to both higher and lower optical frequencies and allows wavelength tunability, but requires at least two well synchronized optical sources at two different optical frequencies. Therefore, each of these interactions alone cannot provide multiple-wavelength or wavelength-tunable output from one, single-wavelength signal source.
Optical parametric interaction is suitable for providing tunable or multiple-wavelength conversion using one, single-wavelength optical signal source. Furthermore, while optical parametric conversion allows conversion of an optical signal only to a lower optical frequency (longer wavelength), by combining parametric interaction with at least one of the above described interactions, any optical frequency above or below the signal-source frequency can be obtained.
The general drawback of parametric optical frequency conversion is that, in order to achieve high parametric gain sufficient to amplify spontaneous quantum-fluctuation noise from microscopic to macroscopic levels and, consequently, to achieve efficient signal-energy conversion, high peak-powers and high pulse-energies are required. It is well known from the prior art that the required energies are well above the energies that can be generated directly from a typical mode-locked, ultrashort-pulse laser oscillator. The best demonstrated result known to date is an optical parametric generation (OPG) threshold at .about.50 nJ, and efficient OPG conversion of .about.40% at approximately 100 nJ achieved in bulk periodically-poled lithium-niobate crystals, as reported by Galvanauskas et al. in "Fiber-laser-based femtosecond parametric generator in bulk periodically poled LiNbO.sub.3 "; Optics Letters, Vol. 22, No. 2; January, 1997. In comparison, typical femtosecond mode-locked pulse energies from a fiber laser are in the range of 10 pJ to 10 nJ (as described by Fermann et al. in "Environmentally stable Kerr-type mode-locked erbium fiber laser producing 360-fs pulses"; Optics Letters; Vol. 19, No. 1; January, 1997, and by Fermann et al. in "Generation of 10 nJ picosecond pulses from a modelocked fibre laser"; Electronics Letters, Vol. 31, No. 3; February, 1995) and those from a solid-state laser are in the range of up to .about.30 nJ (as described by Pelouch et al. in "Ti:sapphire-pumped, high-repetition-rate femtosecond optical parametric oscillator"; Optics Letters, Vol. 17, No. 15; August, 1992).
It is known from the prior art that efficient optical parametric wavelength conversion can be achieved with unamplified or amplified mode-locked laser pulses by arranging a nonlinear crystal in a separate optical cavity in a manner that ensures that pump pulses and signal pulses pass the parametric gain medium synchronously, as seen, for example in the above-referenced article by Pelouch et al. Since, in this case, parametric interaction occurs repetitively, the low, single-pass parametric gain and, consequently, low pulse energies of mode-locked oscillators are sufficient to achieve efficient conversion. The significant practical drawback of this approach is that such a scheme requires two precisely length-matched optical cavities; one for a mode-locked oscillator and another a for synchronously-pumped optical parametric oscillator (OPO). Consequently, such OPO systems are complex, large, and intrinsically very sensitive to the environmental conditions (non-robust). Furthermore, wavelength tuning of such a system requires mechanical movement of the tuning elements such as rotation or translation of a nonlinear crystal, rotation of cavity mirrors, etc., which is incompatible with fast wavelength tuning or switching. Therefore, OPOs can not serve as practical ultrashort-pulse sources for producing multiple-wavelength pulses directly with mode-locked oscillator output.